The current issue and full text archive of this journal is available on Emerald Insight at:

www.emeraldinsight.com/1446-8956.htm

Does PAPI monitoring improve PAPI

monitoring
local governance? Evidence from
a natural experiment in Vietnam
Long Giang and Cuong Nguyen 85
National Economics University, Hanoi, Vietnam and IPAG Lab,
IPAG Business School, Paris, France, and
Anh Tran
Indiana University Bloomington, Bloomington,
Indiana and Harvard University, Cambridge, Massachusetts, USA

Abstract
Purpose – The purpose of this study is to examine the effect of monitoring local authorities on the quality of
governance and public services reported by citizens in Vietnam, using the Vietnam Governance and Public
Administration Performance Index (PAPI) surveys.
Design/methodology/approach – PAPI randomly selected 200 locations in 93 districts of 30 provinces
to conduct its survey in 2010, and subsequently rolled out the survey nationally in 2011 and 2012. Using 2011
and 2012 survey data, the authors compare the quality of governance and public services reported in provinces
and districts that were covered in the 2011 PAPI survey with those that were not surveyed in 2010. Theories
suggest that local authorities may improve their behavior if they have been surveyed and are, thus, aware that
they are being monitored, leading to higher quality governance.
Findings – In this paper, the authors find that governance quality reported in later years by citizens in the
surveyed provinces and districts of the 2010 PAPI survey was significantly higher than the quality reported
by citizens in locations that were not surveyed in 2010. Monitoring appears to improve a wide range of
governance aspects, including local participation in village decisions, transparency of local decision-making,
accountability, administrative procedures and public service delivery.
Originality/value – The main innovation of this study is to use a randomized survey on governance as a
natural experiment to measure the impact of monitoring on the quality of governance and public services, as
reported by citizens.
Keywords Vietnam, Public administration, Governance quality, Governance supervision, PAPI
Paper type Research paper

1. Introduction
As Vietnam’s economy has expanded in recent decades, reaching an annual GDP growth rate
of approximately 7 per cent, not only have living standards improved (Nguyen et al., 2013;
Lanjouw et al., 2016), but there has also been an increasing expectation regarding accessible
and high-quality public services (CECODES et al., 2012). Simultaneously, recognition of the
role of governance and public administration in economic growth and human development is
also increasing in Vietnam (Acuña-Alfaro et al., 2010). Public Administration Reform (PAR),
initiated in Vietnam in the 1990s, continues to drive attempts to improve governance quality
(Painter, 2003; UNDP et al., 2009). The relationship between good governance and economic International Journal of
Development Issues
Vol. 16 No. 1, 2017
JEL classification – H1, H70, G38 pp. 85-106
© Emerald Publishing Limited
The authors are grateful to Jairo Acuña–Alfaro, Do Thi Thanh Huyen (UNDP Vietnam) and Edmund 1446-8956
J. Malesky (Duke University) for their helpful comments on this study. DOI 10.1108/IJDI-05-2016-0028
IJDI growth appears to be mutually reinforcing. In addition to the economic growth in Vietnam
16,1 influencing expectations of good governance, good and effective governance can also
positively influence economic growth and poverty reduction (Aron, 2000; Grindle, 2004;
Martin, 2006; Khan, 2008). For Vietnam’s provinces, the Human Development Index (HDI) is
positively correlated with the Vietnam Governance and Public Administration Performance
Index (PAPI) (UNDP, 2011; CECODES et al., 2012).
86 To better measure and improve governance and public administration in Vietnam, the
United Nations Development Programme (UNDP), together with the Vietnam Fatherland
Front and the Centre for Community Support and Development Studies (CECODES), have in
recent years implemented surveys to collect information for the PAPI. These surveys
measure the respondents’ perceptions of various aspects of their interactions with the local
government. Responses are used to construct an aggregate index of public performance
(CECODES et al., 2012). More specifically, the PAPI survey collected data on citizens’
perception and assessment of the governance quality and public administration of local
authorities on six dimensions: citizen participation at the local level, transparency of
governance, vertical accountability, control of corruption, public administrative procedures
and public-service delivery. The first PAPI was implemented in 2010 in 32 randomly selected
provinces in Vietnam.
A strong body of literature argues that transparency, defined as the ability of the
people to observe and evaluate their governments, allows citizens to hold public agents
accountable, which can reduce corruption and improve public service delivery[1]. When
citizens have better access to information, politicians respond more appropriately to
citizens’ demands, in both developed (Alt et al., 2002) and developing countries (Besley
and Burgess, 2002; Besley et al., 2006; Brunetti and Weder, 2003). However, empirical
findings on the effect of transparency and supervision of authorities have been mixed
(Maskin and Tirole, 2004; Datta, 2008). In our study, we aimed to test whether increased
monitoring leads to improvements in local governance and service quality, as reported
by citizens.
To do so, we approached the first PAPI survey in 2010 as a monitoring intervention on
local governance. This characterization is supported by both central and local government
use of the PAPI data, which is serving as a reference tool for monitoring governance quality
and implementation of PAR in Vietnam (CECODES et al., 2012). Individuals in 93 districts of
30 provinces were sampled in the 2010 PAPI surveys. In 2011 and 2012, the PAPI survey was
expanded to all 63 provinces of Vietnam. Thus, in this study, we used the 2011 and 2012
PAPI surveys as endline surveys to examine whether districts and provinces that were
covered in the 2010 PAPI survey perform better than those that were not covered in the 2010
survey.
The main innovation of this measurement approach is that districts and provinces were
randomly selected into the 2010 PAPI, and as a result, the treatment can be characterized as
a randomized experiment. The randomized design can provide impact estimations with
highly robust internal validity (Duflo, 2006; Duflo et al., 2008; Abhijit et al., 2008). In addition,
the PAPI data are nationally representative, and similar governance surveys have been
implemented in other countries[2], offering robust external validity.
The paper is structured into six sections. The second section presents a brief overview of
related literature. The third and fourth sections present the data sets and estimation
methodology used in this study, respectively. The fifth section discusses the empirical
results from the estimation of the effect of governance monitoring by PAPI surveys on citizen
reports of governance quality and public service delivery. Finally, some conclusions and
policy implications are presented in the sixth section.
2. Literature review PAPI
There is a strong theoretical basis for the argument that an uninformed electorate faces monitoring
obstacles in holding politicians accountable. Buchanan (1989) demonstrated that when
citizens lack information, they are unable to effectively sanction the behavior of politicians,
who can use the opportunity to engage in corruption, self-dealing or catering to more
informed constituents (Besley and Burgess, 2002). Because politicians value holding office, in
environments in which their behavior is being monitored, they have an incentive to serve
their citizens’ interests and behave honestly (Barro, 1973; Ferejohn, 1986). Conversely, in
87
non-transparent environments, the accountability between citizens and their public servants
is much weaker (Alt et al., 2002; Besley and Burgess, 2002; Lassen, 2005).
A number of studies have found empirical support for most of the nodes in the causal
chain linking transparency to improved government behavior. First, there is observational
evidence that well-informed citizens act to hold their officials accountable. More informed
citizens are more likely to pressure officials to act in the interest of their constituents (Lassen,
2005) and to advocate for the fruits of the political process benefiting their localities
(Strömberg, 2004). There is also evidence that politicians respond to greater transparency
with better performance, in both developed (Alt et al., 2002) and developing countries (Besley
and Burgess 2002; Besley et al., 2006; Brunetti and Weder, 2003).
Despite the impressive array of work demonstrating the performance benefits of
transparency, there is by no means consensus on the matter. An alternative literature has
struggled to identify the micro-logic between increased openness, the actions of citizens and
public sector performance (Bauhr and Grimes, 2011; Golden and Picci, 2008). Another set of
scholars argues that increased media attention can damage officials’ performance by
incentivizing politicians to manipulate evaluation, rather than working to enact the most
socially beneficial legislation (Canes-Wrone et al., 2001; Maskin and Tirole, 2004; Datta,
2008).
On the one hand, better information about political behavior may allow media and civil
society to report on political abuses and embarrass political leaders, creating “millions of
auditors” (Kaufmann and Bellver, 2005), as a substitute for direct accountability through
elections (Peruzzotti and Smulovitz, 2006; Smulovitz and Peruzzotti, 2000). This logic
corresponds with the “fire alarm” style of public monitoring first described by McCubbins
and Schwartz (1984). In complex economic and social systems, it can be costly and inefficient
for central government officials to perform police-patrol-style monitoring, in which agents
personally inspect every action taken by subordinate government officials and delegates. An
alternative strategy is to move to a fire alarm approach, in which citizens or media “pull the
alarm” when they see wrongdoing. This system allows the central government to respond to
the abuse without the cost of comprehensive daily inspection. For fire alarm monitoring to be
successful, however, states must maintain a high degree of openness of information so that
citizens and media can play this monitoring role.
On the other hand, however, this approach relies on the untested assumption that
alternative mechanisms can substitute for downward accountability to citizens through
elections (Joshi, 2010). Two variants of this assumption are used by practitioners. A weak
version suggests that public shaming creates a sufficient incentive for the local government
to change its behavior. A stronger version of the assumption is that, once alerted, central
officials will punish self-dealing and corruption. This same logic lies behind the Chinese
Government’s decision to enact its own Open Governance Initiative (OGI) with assistance
from the Yale Law Center (Horsley, 2008; Ma and Wu, 2011). Although the OGI was initiated
by the Chinese state itself and applies to subordinate officials, international donors have also
sought to export the logic of this approach to other contexts, arguing that public shaming or
IJDI the threat of central punishment will incentivize public officials to improve their behavior
16,1 (Joshi, 2010).
A number of studies have examined the effect of monitoring on individuals’ performance.
For example, Duflo et al. (2012) show that class attendance of teachers in India significantly
increases when they are monitored by cameras. Leonard and Masatu (2006) and Leonard
(2008) found that clinicians tend to provide better healthcare services for patients if they are
88 observed by an external research team. However, not all monitoring interventions are
effective in improving the performance of workers (Banerjee and Duflo, 2006).
In Vietnam, the PAPI survey is the first tool for citizens to monitor local governance and
public services, which is provided independently by a civic organization in collaboration
with an international donor. The initial implementation of PAPI was rolled out gradually and
randomly, providing a unique opportunity to evaluate the effect of monitoring on the quality
of governance and services. Our hypothesis is that when local authorities are aware that local
governance and public administration will be measured and evaluated by the PAPI surveys,
they will improve their performance.

3. Data sources
In this study, we draw on data from surveys of PAPI. The PAPI surveys have been
conducted annually by the UNDP, Vietnam Fatherland Front and the CECODES, since 2009
(CECODES et al., 2012). The surveys collect information from citizens regarding many
aspects of their experiences with governance and public administration. Citizens are also
asked about their satisfaction levels with different government organizations at local and
central levels. More detailed descriptions of the PAPI surveys can be found in CECODES
et al. (2012 and 2013).
In 2010, the PAPI survey was conducted in 93 districts of 30 provinces (covering 5,560
citizens). In 2011 and 2012, all 63 provinces were included in the sampling. The 2011 and 2012
survey used the same sample of 207 districts. The PAPI surveys in 2011 and 2012 covered
13,642 and 13,747 respondents, respectively. Across the three surveys (2010, 2011 and 2012),
there were 40 districts that were sampled in all three years.

4. Methodology
There are a large number of interventions on governance and public administration reforms
ongoing in Vietnam, and to our knowledge there have been no randomized experiments on
any of these interventions[3]. When a policy or program is not randomly assigned, the impact
of the policy or program is difficult to estimate due to selection bias. Recently, Malesky et al.
(2014) examined a quasi-experimental intervention of removal of district-level People’s
Councils in Vietnam, but this intervention lacked randomized selection.
In the present study, rather than measuring the effect of a central-government-initiated
intervention to monitor local governance (which the government would be unlikely to
conduct randomly), we instead approached the 2010 PAPI survey as a type of randomized
intervention. Although the survey is administered through non-governmental actors, there is
evidence that Vietnam’s Central Government has been using PAPI as a reference tool in
monitoring governance quality and PAR (CECODES et al., 2012). At the provincial level,
several provinces such as Kon Tum, Da Nang City and Ho Chi Minh City also use information
from PAPI to guide policies on public governance improvement (CECODES et al., 2012).
Figure 1 depicts the study design. Our treatment is the 2010 PAPI survey, which is a
proxy of monitoring and supervision of governance and public administration at the local
areas. Among the 63 cities and provinces of Vietnam, 30 cities and provinces were selected in
the 2010 PAPI survey. The Ha Noi and Ho Chi Minh cities were selected purposely, and 28
provinces were randomly selected. All 87 districts in these 28 provinces were selected for the
 PAPI
All 63 provinces
and cities monitoring

30 provinces and Control group: 33

cities covered in the provinces and cities
not covered in the
2010 PAPI survey
2010 PAPI survey 89

Treatment group: 28
Hanoi capita city and provinces and cities;
HCM city 87 districts in these
provinces

47 districts that Treatment group: Control group: 167

are not sampled in 40 districts that are districts that are
the 2011 and 2012 also sampled in the sampled in the 2011
PAPI surveys 2011 and 2012 PAPI and 2012 PAPI
Figure 1.
Treatment and control
Source: Authors’ preparation group in this study

2010 PAPI surveys. Of 87 districts, 40 districts were further selected in the 2011 and 2012
PAPI surveys. In this study, we compared the quality of governance and public
administration in 2012 of districts in the 2010 PAPI survey with those of the districts not in
the 2010 PAPI survey. Thus, the treatment group consists of districts and provinces that are
sampled in the 2010 PAPI survey and the 2012 PAPI survey, while the control group consists
of districts and provinces that are sampled in the 2012 PAPI survey but not in the 2010 PAPI
survey.
Leaders within districts and provinces that are sampled for PAPI are undoubtedly aware
of the survey. Local officials can be expected to recognize that citizens have reported and will
continue to report on their experiences with local governance and administration. This
awareness may influence leaders to improve their performance. To test this hypothesis, we
measured the effect of districts being randomly selected into the 2010 PAPI survey on both
governance and public-service delivery outcomes as reported by citizens.
The main advantage of the above approach is that provinces and districts are randomly
selected into the PAPI. As a result, the treatment can be considered as a randomized
experiment. As known, the randomized design is an emerging method which can provide the
ideal estimator of impact evaluation with robust internal validity (Duflo, 2006; Duflo et al.,
2008; Abhijit et al., 2008). In addition, the PAPI data are national representative, and similar
governance surveys are implemented in other countries[4]. The impact evaluation, therefore,
can provide good external validity.
In this study, we examined the effect of the PAPI surveys on different citizen-reported
outcomes, including governance and public administration and public services. First, we
measured the effect of the selection of provinces in the 2010 PAPI survey on the governance
quality in the 2012 PAPI survey as follows:

Yip, 2012 ⫽ ␣ ⫹ ␤PAPIp, 2010 ⫹ Xip␥ ⫹ ␧ip (1)

IJDI where Yip, 2012 is an indicator of governance quality of individual i in province p, measured in
16,1 the 2012 PAPI survey. PAPIp, 2010 is the dummy variable, indicating whether province p was
also surveyed in the 2010 PAPI survey. Xid is a vector of control variables.
Second, we measured the effect of the selection of districts in the 2010 PAPI survey on the
governance quality in the 2012 PAPI survey as follows:

90 Yid, 2012 ⫽ ␣ ⫹ ␤PAPId, 2010 ⫹ Xid␥ ⫹ ␧id (2)

where Yid, 2012 is an indicator of governance quality of individual i in district d, measured in

the 2012 PAPI survey. PAPId, 2010 is the dummy variable, indicating whether district d was
surveyed in the 2010 PAPI survey. Xid is a vector of control variables. For both Models (1)
and (2), we used a small set of exogenous control variables including the population of
provinces and the population of districts. For Model 2, we included a dummy variable
indicating whether the headquarters for the province is located in the surveyed district.
We also ran regressions of similar specifications as equations (1) and (2) with the
dependent variables of the 2011 PAPI to examine whether there is an effect of the 2010 PAPI
on the respondents in the 2011 PAPI.
There are two points that should be noted in the above models. First, in the 2010 PAPI
surveys, although 28 provinces were randomly selected, Ha Noi City and the Ho Chi Minh
City were deliberately selected. To maintain a sample of only randomly selected provinces,
we dropped households from Ha Noi City and the Ho Chi Minh City from the sample. Second,
in all three years of the PAPI surveys, urban districts with headquarters of provinces were
always selected. It means that the treatment group was urban districts. Thus, we dropped all
rural households from regressions. The number of households used in regressions in
equation (1) and (2) was 7,597 individuals.
It is important to ascertain whether the PAPI provinces and districts were indeed
randomly sampled in the 2010 PAPI survey. CECODES et al. (2011) conducted a balancing
test to test the differences in outcomes between provinces that were sampled in the 2010
PAPI survey and provinces that were not sampled in the 2010 PAPI survey. The outcome
means of the sampled and non-sampled provinces were similar. In this study, we also tested
whether the characteristics of households in the sampled provinces and households in the
non-sampled provinces were statistically different. We used the Vietnam Household Living
Standard Survey 2010. Tables AI and AII present the regressions. In all regressions, there
were no statistically significant effects of the 2010 PAPI survey on household characteristics.

5. Empirical results
Measuring governance quality and public administration is challenging, as governance and
public administration are multidimensional. In this study, we followed the approach of
composite indexes in CECODES et al. (2012 and 2013), which measure governance quality
and public administration along six dimensions: citizen participation at the local level,
transparency of governance, vertical accountability, control of corruption, public
administrative procedures and public-service delivery. Each dimension comprises several
sub-dimensions. The sub-dimensions are also measured by composite indexes, which are
computed from several indicators. The six dimensions are measured on a scale from 1 to 10,
with 1 indicating the poorest performance and 10 indicating the best performance of
governance and public administration. A full description of the aggregate indexes of
dimensions is presented in CECODES et al. (2012 and 2013).
Based on the six dimensions, CECODES et al. (2012 and 2013) constructed an overall
composite measure of governance and public administration, which is the PAPI. This index
is the weighted sum (or unweighted sum) of the indexes of the six dimensions. PAPI scores PAPI
vary from 6 (the lowest possible score) to 60 (the highest possible score). monitoring
Table I reports regressions in the 2012 PAPI and the indexes of six dimensions on the
dummy variable indicating whether provinces were surveyed in the 2010 PAPI. This table
shows the effect of provinces being surveyed in the 2010 PAPI on survey respondents’
perceptions of governance quality reported in the 2012 PAPI. There are no control variables
in the regressions in Table I. As mentioned above, the sample for regression includes only
urban individuals and there are no individuals from Ha Noi City and Ho Chi Minh City. 91
Interestingly, the effect of the 2010 PAPI survey is positive and statistically significant in all
dimensions. The weighted PAPI increases by 0.8. Because the average weighted PAPI is 39,
the effect of the 2010 PAPI survey on the weighted PAPI is approximately 2.1 per cent. The
effect of the unweighted PAPI is also positive and statistically significant. Although the
effect is relatively small, it is significant. These findings indicate that when local authorities
are aware that their performance is being assessed and evaluated, their performance tends to
improve in the eyes of their constituents.
In Table II, several control variables are added in the regressions. The effect of the 2010
PAPI survey on the governance indexes in the 2011 PAPI is also positive and statistically
significant in all the dimensions, with the exception of the dimension “control of corruption”.
Tables III and IV present the regression of respondents’ satisfaction with governance.
The contains data on citizen satisfaction with respect to different levels of government.
Again, people in provinces that were surveyed in the 2010 PAPI are more likely to report
higher satisfaction with governance than those in provinces not sampled in the 2010 PAPI.
These findings imply that using governance surveys as a monitoring tool of local
governance can help improve satisfaction with governance quality and public
administration procedures.
Tables V and VI present the regression of the governance indexes and citizen satisfaction
on the dummy variable of districts being sampled in the 2010 PAPI survey. Similarly, the
effect is positive and statistically significant in the regression of the PAPI indexes. However,
for the dimensions of “accountability” and “control of corruption”, the effect of districts being
sampled in the 2010 PAPI survey is not statically significant. Regarding the overall
satisfaction of citizens with governance quality, the effect of districts being sampled in the
2010 PAPI also had a significant and positive effect in most aspects.
Finally, Tables AI–AVI present the replicated estimation of the effect of the 2010 PAPI on
the 2011 PAPI indexes. Overall, we found a positive effect of the 2010 PAPI on citizens’
reports of governance and public administration quality in 2011.

6. Conclusion
Using the PAPI data, this study examined the effect of monitoring of local governance on
citizen reports of governance quality and public service delivery with two main research
questions:
RQ1. To what extent does monitoring of local governance by PAPI strengthen local
governance as reported by citizens?
RQ2. To what extent does the monitoring of local governance by PAPI improve the
quality of the public service delivery as reported by citizens?
We found that the 2010 PAPI survey has a positive and statistically significant effect on
citizens’ perceptions of public administration and governance in both 2011 and 2012. The
positive effects are found in most dimensions of governance and public administration,
except for the dimension of control of corruption. The weighted PAPI and unweighted PAPI
 92
IJDI
16,1

Table I.

variables
the 2010 PAPI:
without control
2012 PAPI on the
Regressions of the

provinces covered in
Dimension 1: Dimension 2:
Quality of Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted participation in of local Downward Control of Administrative Public service
variables PAPI in 2012 PAPI in 2012 village decisions decision-making accountability corruption procedures delivery

Provinces covered 0.811*** (0.054) 0.807*** (0.054) 0.064*** (0.016) 0.169*** (0.016) 0.086*** (0.013) 0.075*** (0.025) 0.278*** (0.011) 0.135*** (0.007)
in PAPI 2010
Constant 38.910*** (0.038) 36.749*** (0.037) 5.202*** (0.010) 5.850*** (0.010) 5.651*** (0.008) 5.867*** (0.018) 6.936*** (0.007) 7.242*** (0.005)
Observations 7,697 7,697 7,697 7,697 7,697 7,697 7,697 7,697
R2 0.007 0.007 0.001 0.003 0.001 0.000 0.024 0.010

Notes: Standard errors in parentheses; *significant at 10%; **significant at 5%; *** significant at 1%
Source: Estimation from the 2012 PAPI survey
 Dimension 1: Dimension 2:
Quality of Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted participation in of local Downward Control of Administrative Public service
variables PAPI in 2012 PAPI in 2012 village decisions decision-making accountability corruption procedures delivery

Provinces covered
in PAPI 2010 0.625*** (0.057) 0.618*** (0.057) 0.031* (0.017) 0.129*** (0.017) 0.062*** (0.014) 0.042 (0.026) 0.244*** (0.011) 0.110*** (0.007)
Provincial People’s
Committee is located
in districts 0.525*** (0.059) 0.368*** (0.058) 0.123*** (0.017) 0.194*** (0.017) 0.001 (0.014) -0.120*** (0.028) 0.060*** (0.011) 0.110*** (0.008)
Total population of
district in 2009 ⫺1.401*** (0.363) ⫺3.136*** (0.390) ⫺0.646*** (0.145) ⫺0.825*** (0.104) ⫺0.898*** (0.085) ⫺0.762*** (0.158) ⫺0.433*** (0.082) 0.428*** (0.053)
Total population of
province in 2009 ⫺0.663*** (0.045) ⫺0.716*** (0.045) ⫺0.152*** (0.013) ⫺0.165*** (0.013) ⫺0.107*** (0.011) ⫺0.245*** (0.022) ⫺0.017** (0.009) ⫺0.030*** (0.006)
Large provinces/cities
(yes ⫽ 1, no ⫽ 0) 0.755*** (0.089) 0.802*** (0.090) 0.109*** (0.025) 0.068** (0.028) 0.086*** (0.025) 0.133*** (0.039) 0.257*** (0.018) 0.149*** (0.011)
Constant 39.706*** (0.075) 37.956*** (0.078) 5.439*** (0.025) 6.107*** (0.022) 5.924*** (0.019) 6.358*** (0.034) 6.980*** (0.015) 7.148*** (0.010)
Observations 7,594 7,594 7,594 7,594 7,594 7,594 7,594 7,594
R2 0.023 0.028 0.014 0.018 0.014 0.012 0.036 0.035

Notes: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2012 PAPI survey
93
PAPI

control variables
Regressions of the
Table II.
monitoring

the 2010 PAPI: with

provinces covered in
2012 PAPI on the
 94
IJDI
16,1

variables
Table III.

without control
2012 satisfaction

in the 2010 PAPI:

Regressions of the

the provinces covered

about governances on
Satisfaction Satisfaction
Satisfaction about Commune about District Satisfaction about Your satisfaction Your satisfaction Satisfaction
Explanatory about village People’s People’s Province People’s Satisfaction about National about commune about district
variables head Committee Committee Committee about Government Assembly security court

Provinces covered
in PAPI 2010 1.490*** (0.203) 1.788*** (0.215) 1.838*** (0.264) 1.944*** (0.313) 0.750** (0.311) 1.374*** (0.296) 1.956*** (0.250) 1.229*** (0.310)
Constant 86.718*** (0.149) 83.941*** (0.155) 84.567*** (0.175) 85.428*** (0.190) 88.164*** (0.178) 88.506*** (0.177) 83.592*** (0.164) 85.266*** (0.177)
Observations 7,261 6,572 4,738 4,055 4,275 4,231 6,348 3,951
R2 0.002 0.003 0.003 0.003 0.000 0.002 0.003 0.001

Notes: Standard errors in parentheses; *significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2012 PAPI survey
 Satisfaction about
Commune Satisfaction about Satisfaction about Your satisfaction
Explanatory Satisfaction about People’s District People’s Province People’s Satisfaction about Your satisfaction about about commune Satisfaction about
variables village head Committee Committee Committee Government National Assembly security district court

Provinces covered
in PAPI 2010 1.190*** (0.209) 1.560*** (0.226) 1.532*** (0.285) 1.810*** (0.338) 0.922*** (0.337) 1.566*** (0.321) 1.431*** (0.261) 0.991*** (0.317)
Provincial People’s
Committee is
located in districts 1.389*** (0.229) 0.792*** (0.235) ⫺0.161 (0.285) ⫺0.598* (0.313) ⫺0.791*** (0.283) 0.242 (0.268) 2.300*** (0.252) ⫺0.026 (0.307)
Total population of
district in 2009 ⫺5.002*** (1.442) ⫺1.905 (1.542) ⫺0.367 (1.902) ⫺9.176** (3.723) ⫺12.246*** (4.397) ⫺13.908*** (4.250) ⫺7.422*** (2.202) ⫺7.981** (3.093)
Total population of
province in 2009 ⫺0.217 (0.164) ⫺0.798*** (0.185) ⫺0.630*** (0.196) 0.289 (0.219) ⫺0.915*** (0.205) ⫺0.261 (0.199) ⫺0.834*** (0.195) ⫺0.440* (0.223)
Large provinces/cities
(yes ⫽ 1, no ⫽ 0) 1.932*** (0.268) 0.397 (0.311) 1.255*** (0.426) 1.813*** (0.444) ⫺1.352*** (0.454) ⫺1.223*** (0.442) 1.938*** (0.359) 1.440*** (0.469)
Constant 87.011*** (0.293) 84.919*** (0.306) 85.469*** (0.377) 86.681*** (0.566) 91.803*** (0.634) 91.069*** (0.599) 84.824*** (0.396) 87.077*** (0.530)
Observations 7,161 6,479 4,664 3,995 4,210 4,167 6,255 3,889
R2 0.006 0.004 0.004 0.008 0.014 0.014 0.010 0.005

Notes: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2012 PAPI survey
95

control variables
Table IV.
monitoring
PAPI

in the 2010 PAPI: with

the provinces covered
about governances on
2012 satisfaction
Regressions of the
 96
IJDI
16,1

Table V.

2010 PAPI: with

control variables
2012 PAPI on the
Regressions of the

districts covered in the

Dimension 1: Dimension 2:
Quality of Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted participation in of local Downward Control of Administrative Public service
variables PAPI in 2012 PAPI in 2012 village decisions decision-making accountability corruption procedures delivery

Districts covered
in PAPI 2010 0.440*** (0.074) 0.314*** (0.074) ⫺0.041* (0.022) 0.087*** (0.022) 0.001 (0.017) ⫺0.059 (0.035) 0.206*** (0.014) 0.121*** (0.010)
Provincial People’s
Committee is located
in districts 0.388*** (0.066) 0.277*** (0.065) 0.139*** (0.019) 0.167*** (0.019) 0.003 (0.015) ⫺0.097*** (0.031) ⫺0.006 (0.012) 0.070*** (0.008)
Total population of
district in 2009 ⫺1.540*** (0.362) ⫺3.231*** (0.387) ⫺0.631*** (0.143) ⫺0.852*** (0.103) ⫺0.897*** (0.084) ⫺0.741*** (0.157) ⫺0.499*** (0.081) 0.389*** (0.053)
Total population of
province in 2009 ⫺0.676*** (0.045) ⫺0.731*** (0.045) ⫺0.154*** (0.013) ⫺0.168*** (0.013) ⫺0.109*** (0.011) ⫺0.247*** (0.022) ⫺0.022** (0.009) ⫺0.031*** (0.006)
Large provinces/cities
(yes ⫽ 1, no ⫽ 0) 0.936*** (0.085) 0.972*** (0.086) 0.113*** (0.024) 0.106*** (0.027) 0.100*** (0.024) 0.138*** (0.038) 0.331*** (0.018) 0.184*** (0.011)
Constant 39.955*** (0.068) 38.210*** (0.071) 5.455*** (0.022) 6.159*** (0.020) 5.952*** (0.018) 6.381*** (0.031) 7.075*** (0.014) 7.189*** (0.009)
Observations 7,594 7,594 7,594 7,594 7,594 7,594 7,594 7,594
R2 0.020 0.025 0.014 0.017 0.014 0.012 0.027 0.033

Notes: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
 Satisfaction about Satisfaction about Satisfaction about Your satisfaction
Explanatory Satisfaction about Commune People’s District People’s Province People’s Satisfaction about Your satisfaction about about commune Satisfaction about
variables village head Committee Committee Committee Government National Assembly security district court

Districts covered
in PAPI 2010 0.774*** (0.269) 1.097*** (0.287) 0.560 (0.365) 0.496 (0.397) 0.889** (0.387) 1.298*** (0.369) 1.010*** (0.311) 0.381 (0.409)
Provincial People’s
Committee is
located in districts 1.142*** (0.255) 0.418 (0.267) ⫺0.339 (0.303) ⫺0.755** (0.339) ⫺1.110*** (0.325) ⫺0.221 (0.311) 1.965*** (0.265) ⫺0.128 (0.315)
Total population of
district in 2009 ⫺5.258*** (1.432) ⫺2.194 (1.532) ⫺0.518 (1.884) ⫺9.165** (3.678) ⫺12.549*** (4.338) ⫺14.312*** (4.200) ⫺7.721*** (2.179) ⫺8.052** (3.046)
Total population of
province in 2009 ⫺0.238 (0.164) ⫺0.836*** (0.184) ⫺0.688*** (0.195) 0.191 (0.219) ⫺0.918*** (0.203) ⫺0.281 (0.198) ⫺0.868*** (0.194) ⫺0.473** (0.222)
Large provinces/cities
(yes ⫽ 1, no ⫽ 0) 2.273*** (0.262) 0.857*** (0.301) 1.699*** (0.397) 2.323*** (0.401) ⫺1.037** (0.394) ⫺0.695* (0.386) 2.385*** (0.335) 1.719*** (0.447)
Constant 87.497*** (0.254) 85.572*** (0.270) 86.159*** (0.325) 87.527*** (0.473) 92.170*** (0.535) 91.714*** (0.504) 85.419*** (0.342) 87.501*** (0.455)
Observations 7,161 6,479 4,664 3,995 4,210 4,167 6,255 3,889
R2 0.005 0.003 0.002 0.006 0.014 0.012 0.009 0.004

Notes: Standard errors in parentheses; * significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
97

control variables
Table VI.
monitoring
PAPI

the 2010 PAPI: with

the districts covered in
about governances on
2012 satisfaction
Regressions of the
IJDI in 2012 were also increased by approximately 2.1 per cent by the 2010 PAPI survey. Citizens’
16,1 reported satisfaction with local and central governance also increased. These findings imply
that using governance surveys as a monitoring tool can help improve local governance
quality and public administration. However, a limitation of our study is that the
measurement tool is also the intervention. It is possible that the opinions of respondents in
repeatedly surveyed districts were positively influenced by the surveyors’ interest in seeking
98 those opinions. Although the survey was conducted by non-governmental actors, residents
may perceive the survey as a sign that the government is seeking citizen input, which could
lead to higher regard for the government. Future studies should evaluate government
performance using empirical measures external to the survey.
The findings in this paper have several policy implications. First, it is worthwhile to
consider expanding the geographical coverage of PAPI. As a monitoring tool, the current
small sample would be sufficient. However, given the evidence that PAPI is effective in
improving local governance, it is desirable to expand PAPI’s coverage to more districts and
communes. Second, policymakers should consider exploring approaches similar to PAPI in
other public arenas. Ministries and Government agencies in Vietnam provide a wide range of
public services to citizens. Having a PAPI-style tool to improve transparency of public
services provided by ministries and government agencies could prove beneficial. Finally, the
results of this study could also be used to guide transparency programs in other countries
that are working to improve local governance quality.

Notes
1. This literature review is based on Malesky et al. (2012).
2. Examples include Indonesia Governance and Decentralization Survey, Pakistan Social Audit of
Local Governance Survey, Cambodia Governance and Corruption Diagnostic, Households and
Local Self-governance in Russia Regions, World Bank multiple-country Governance Diagnostic
Study.
3. For review of impact evaluation of programs and policies in Vietnam, see Nguyen (2015).
4. For example, China has a similar survey on governance (Saich, 2007).

References
Abhijit, V., Banerjee, A. and Duflo, E. (2008), “The experimental approach to development economics”,
NBER Working Paper No. 14467, National Bureau of Economic Research, Cambridge.
Acuña-Alfaro, J., Do, G. and Do, H. (2010), “Measuring governance and public administration for human
development: a demand-side approach”, Background Paper for 2010 National Human
Development Report, UNDP and VASS, Hanoi.
Alt, J., Lassen, D. and Skilling, D. (2002), “Fiscal transparency, gubernatorial approval, and the scale of
government: evidence from the States”, State Politics & Policy Quarterly, Vol. 2 No. 3, pp. 230-250.
Aron, J. (2000), “Growth and institutions: a review of evidence”, The World Bank Research Observer,
Vol. 15 No. 1, pp. 99-135.
Banerjee, A. and Duflo, E. (2006), “Addressing absence”, Journal of Economic Perspectives, Vol. 20 No. 1,
pp. 117-132.
Barro, R. (1973), “The control of politicians: an economic model”, Public Choice, Vol. 14 No. 1, pp. 19-42.
Bauhr, M. and Grimes, M. (2011), “Seeing the state: the implications of transparency for societal
accountability”, presented at the Annual Meeting of the Midwest Political Science Conference,
Chicago, IL.
Besley, T. and Burgess, R. (2002), “The political economy of government responsiveness: theory and
evidence from India”, Quarterly Journal of Economics, Vol. 117, November, pp. 1415-1452.
Besley, T., Pande, R. and Rao, V. (2006), “Political selection and the quality of government: evidence PAPI
from South India”, Unpublished manuscript, London School of Economics.
monitoring
Brunetti, A. and Weder, B. (2003), “A free press is bad news for corruption”, Journal of Public Economics,
Vol. 87 Nos 7/8, pp. 1801-1824.
Buchanan, J.M. (1989), “The public-choice perspective”, Essays on the Political Economy, University of
Hawaii Press, Honolulu, pp. 13-24.
Canes-Wrone, B., Herron, M. and Shotts, K. (2001), “Leadership and pandering: a theory of executive 99
policymaking”, American Journal of Political Science, Vol. 45 No. 3, pp. 532-550.
CECODES (Centre for Community Support and Development Studies), FR (The Front Review of the
Central Committee for the Vietnam’s Fatherland Front) and UNDP (United Nations Development
Programme) (2011), The Viet Nam Governance and Public Administration Performance Index
(PAPI): Measuring Citizens’ Experiences, UNDP Viet Nam, Hanoi.
CECODES, FR, CPP (Commission on People’s Petitions of the Standing Committee for the National
Assembly of Viet Nam) and UNDP (2012), The Viet Nam Governance and Public Administration
Performance Index (PAPI): Measuring Citizens’ Experiences, UNDP Viet Nam, Hanoi.
CECODES, VFF-CRT (Centre for Research and Training of the Vietnam Fatherland Front) and UNDP
(2013), The Vietnam Governance and Public Administration Performance Index (PAPI) 2012:
Measuring Citizens’ Experiences, UNDP Viet Nam, Hanoi.
Datta, S. (2008), “Television coverage and political voice: evidence from parliamentary question hour in
India”, available at: http://papers.ssrn.com/sol3/papers.cfm?abstract_id⫽1281627
Duflo, E. (2006), “Field experiments in development economics”, prepared for the World Congress of the
Econometric Society, available at: http://economics.mit.edu/files/800
Duflo, E., Glennerster, R. and Kremer, M. (2008), “Using randomization in development economics
research: a toolkit”, Handbook of Development Economics, Vol. 4.
Duflo, E., Hanna, R. and Ryan, S. (2012), “Incentives work: getting teachers to come to school”, American
Economic Review, Vol. 102 No. 4, pp. 1241-1278.
Ferejohn, J. (1986), “Incumbent performance and electoral control”, Public Choice, Vol. 50 Nos 1/2/3,
pp. 5-25.
Golden, M. and Picci, L. (2008), “Pork barrel politics in postwar Italy, 1953-1992”, American Journal of
Political Science, Vol. 52 No. 2, pp. 268-289.
Grindle, M.S. (2004), “Good enough governance: poverty reduction and reform in developing countries”,
Governance, Vol. 17 No. 4, pp. 525-548.
Horsley, J. (2008), “China adopts first nationwide open government information regulations”, Yale China
Law Center Working Paper, available at: www.law.yale.edu/documents/pdf/Intellectual_Life/
Ch_China_Adopts_1st_OGI_Regulations.pdf
Joshi, A. (2010), “Annex 1: service delivery”, in McGee, R. and Gaventa, J. (Eds), Review of Impact and
Effectiveness of Transparency and Accountability Initiatives: A Review of the Evidence to Date,
Institute of Development Studies, Sussex.
Kaufmann, D. and Bellver, A. (2005), “Transplanting transparency: initial empirics and policy
applications”, World Bank Policy Research Working Paper, available at: http://ssrn.com/
abstract⫽808664orhttp://dx.doi.org/10.2139/ssrn.808664
Khan, M.H. (2008), “Governance and development: the perspective of growth-enhancing governance”, in
GRIPS Development Forum (Ed.), Diversity and Complementarity in Development Aid: East
Asian Lessons for African Growth, National Graduate Institute for Policy Studies, Tokyo.
Lanjouw, P., Marra, M. and Nguyen, V.C. (2016), “Vietnam’s evolving poverty index map: patterns and
implications for policy”, Social Indicators Research, doi: 10.1007/s11205-016-1355-9.
Lassen, D. (2005), “The effect of information on voter turnout: evidence from a natural experiment”,
American Journal of Political Science, Vol. 49 No. 1, pp. 103-118.
IJDI Leonard, K.L. (2008), “Is patient satisfaction sensitive to changes in the quality of care? An exploitation
of the Hawthorne effect”, Journal of Health Economics, Vol. 27 No. 2, pp. 444-459.
16,1
Leonard, K.L. and Masatu, M.C. (2006), “Outpatient process quality evaluation and the hawthorne
effect”, Social Science & Medicine, Vol. 63 No. 9, pp. 2330-2340.
McCubbins, M.D. and Schwartz, T. (1984), “Congressional oversight overlooked: policy patrols versus
fire alarms”, American Journal of Political Science, Vol. 28 No. 1, pp. 165-179.
100 Ma, L. and Wu, J. (2011), “What drives fiscal transparency? Evidence from provincial governments in
China”, presented at the 1st Global Conference on Transparency Research, Rutgers University,
Newark, NJ.
Malesky, E., Nguyen, V.C. and Tran, A. (2014), “The impact of recentralization on public services: a
quasi-experiment on abolishing elected councils in Vietnam”, American Political Science Review,
Vol. 108 No. 1, pp. 144-168.
Malesky, E., Schuler, P. and Tran, A. (2012), “The adverse effects of sunshine: a field experiment on
legislative transparency in an authoritarian assembly”, American Political Science Review,
Vol. 106 No. 4, pp. 762-786, doi:10.1017/S0003055412000.
Martin, M. (2006), “Governance and poverty reduction: paths of connection”, Trocaire Development
Review, pp. 35-55.
Maskin, E. and Tirole, J. (2004), “The politician and the judge: accountability in government”, The
American Economic Review, Vol. 94 No. 4, pp. 1034-1054.
Nguyen, C. (2015), “Impact evaluation of development programmes and policies: experiences from
Vietnam”, Working Papers 2015-620, Department of Research, IPAG Business School, Paris.
Nguyen, C., Linh, V. and Nguyen, T. (2013), “Urban poverty in Vietnam: determinants and policy
implications”, International Journal of Development Issues, Vol. 12 No. 2, pp. 110-139.
Painter, M. (2003), “Public administration reform in Vietnam: problems and prospects”, Public
Administration and Development, Vol. 23 No. 3, pp. 259-271.
Peruzzotti, E. and Smulovitz, C. (Eds) (2006), Enforcing The Rule of Law: Social Accountability in The
New Latin American Democracies, University of Pittsburgh Pre.
Saich, T. (2007), “Citizens’ perceptions of governance in rural and urban China”, Journal of Chinese
Political Science, Vol. 12 No. 1, doi: 10.1007/s11366-007-9003-5.
Smulovitz, C. and Peruzzotti, E. (2000), “Societal accountability in Latin America”, Journal of
Democracy, Vol. 1 No. 4, pp. 147-158.
Strömberg, D. (2004), “Radio’s impact on public spending”, Quarterly Journal of Economics, Vol. 119
No. 1, pp. 189-212.
UNDP (United Nations Development Programme) (2011), Social Service for Human Development:
Vietnam Human Development Report 2011, UNDP Viet Nam, Hanoi.
UNDP, VFF (Vietnam’s Fatherland Front) and CECODES (2009), Reforming Public Administration in
Viet Nam: Current Situation and Recommendations, National Political Publishing House, Hanoi.

Corresponding author
Cuong Nguyen can be contacted at: [email protected]
 Appendix

Dimension 1: Dimension 2:
Quality of Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted participation in of local Downward Control of Administrative Public service
variables PAPI in 2011 PAPI in 2011 village decisions decision-making accountability corruption procedures delivery

Provinces covered
in PAPI 2010 0.445*** (0.038) 0.548*** (0.038) 0.093*** (0.010) 0.101*** (0.013) 0.141*** (0.011) ⫺0.013(0.014) 0.151*** (0.008) 0.075*** (0.005)
Constant 38.382*** (0.026) 36.773*** (0.026) 5.267*** (0.007) 5.640*** (0.008) 5.534*** (0.007) 6.334*** (0.010) 6.930*** (0.005) 7.069*** (0.004)
Observations 7,597 7,597 7,597 7,597 7,597 7,597 7,597 7,597
R2 0.002 0.004 0.001 0.001 0.003 0.000 0.007 0.003

Notes: Standard errors in parentheses; *significant at 10%; **significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey

control variables
Table AI.
101
monitoring
PAPI

PAPI 2010: without

provinces covered in
2011 PAPI on
Regressions of the
 IJDI
16,1

102

Table AII.

2011 PAPI on

PAPI 2010: with

control variables
Regressions of the

provinces covered in
Dimension 1:
Quality of Dimension 2:
participation Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted PAPI in village of local Downward Control of Administrative Public service
variables PAPI in 2011 in 2011 decisions decision-making accountability corruption procedures delivery

Provinces covered
in PAPI 2010 0.449*** (0.040) 0.572*** (0.040) 0.094*** (0.011) 0.096*** (0.013) 0.145*** (0.011) 0.016 (0.014) 0.157*** (0.008) 0.064*** (0.006)
Provincial People’s
Committee is located
in districts 1.078*** (0.043) 0.796*** (0.043) 0.192*** (0.012) 0.259*** (0.013) 0.143*** (0.012) ⫺0.052*** (0.016) ⫺0.023** (0.009) 0.278*** (0.006)
Total population of
district in 2009 0.258 (0.201) 0.629*** (0.200) 0.026 (0.049) ⫺0.186*** (0.063) ⫺0.168*** (0.058) 1.281*** (0.080) 0.058 (0.048) ⫺0.383*** (0.030)
Total population of
province in 2009 ⫺0.027(0.033) ⫺0.278*** (0.033) ⫺0.060*** (0.009) ⫺0.108*** (0.010) ⫺0.084*** (0.009) ⫺0.071*** (0.012) ⫺0.083*** (0.007) 0.128*** (0.005)
City class 1 ⫺0.399*** (0.056) ⫺0.618*** (0.058) ⫺0.079*** (0.016) ⫺0.112*** (0.020) ⫺0.073*** (0.015) ⫺0.403*** (0.020) ⫺0.086*** (0.013) 0.134*** (0.007)
Constant 37.916*** (0.047) 36.718*** (0.047) 5.257*** (0.013) 5.703*** (0.016) 5.607*** (0.014) 6.279*** (0.017) 7.047*** (0.010) 6.825*** (0.007)
Observations 7,507 7,507 7,507 7,507 7,507 7,507 7,507 7,507
R2 0.018 0.014 0.008 0.011 0.008 0.014 0.010 0.049

Notes: Standard errors in parentheses; *significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
 Satisfaction about
Commune Satisfaction about Satisfaction about Your satisfaction Your satisfaction
Explanatory Satisfaction about People’s District People’s Province People’s Satisfaction about about National about commune Satisfaction about
variables village head Committee Committee Committee Government Assembly security district court

Provinces covered 3.005*** (0.195) 1.440*** (0.205) 1.213*** (0.212) 1.585*** (0.246) 1.497*** (0.231) 1.620*** (0.225) 2.603*** (0.215) 2.339*** (0.276)
in PAPI 2010
Constant 83.583*** (0.121) 82.621*** (0.120) 83.566*** (0.136) 85.060*** (0.164) 89.511*** (0.147) 90.174*** (0.145) 81.600*** (0.158) 83.309*** (0.178)
Observations 7,050 6,226 4,221 3,472 3,725 3,696 5,999 3,412
R2 0.006 0.001 0.001 0.002 0.002 0.003 0.004 0.003

Notes: Standard errors in parentheses; *significant at 10%; **significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
103

control variables
Regressions of the
Table AIII.
monitoring
PAPI

PAPI 2010: without

provinces covered in
about governance on
2011 satisfaction
 IJDI
16,1

104

Table AIV.

PAPI 2010: with

control variables
2011 satisfaction
Regressions of the

provinces covered in
about governance on
Satisfaction about
Commune Satisfaction about Satisfaction about Your satisfaction Your satisfaction
Explanatory Satisfaction about People’s District People’s Province People’s Satisfaction about about National about commune Satisfaction about
variables village head Committee Committee Committee Government Assembly security district court

Provinces covered
in PAPI 2010 2.946*** (0.199) 1.433*** (0.210) 1.548*** (0.218) 2.190*** (0.250) 1.975*** (0.242) 1.944*** (0.234) 2.675*** (0.226) 2.762*** (0.302)
Provincial People’s
Committee is
located in districts 1.990*** (0.207) 1.588*** (0.224) ⫺0.088 (0.233) ⫺1.690*** (0.251) ⫺1.511*** (0.235) ⫺1.091*** (0.234) 2.290*** (0.229) 1.003*** (0.281)
Total population of
district in 2009 ⫺2.963** (1.373) ⫺4.772*** (0.821) 7.057*** (0.912) ⫺1.414 (1.053) ⫺2.211** (0.932) 2.984*** (0.977) ⫺3.134*** (1.049) 2.530 (1.659)
Total population of
province in 2009 ⫺0.357** (0.161) 0.637*** (0.170) ⫺0.269 (0.189) 1.531*** (0.218) 0.942*** (0.201) 0.594*** (0.196) 0.694*** (0.194) 1.197*** (0.227)
City class 1 ⫺0.026 (0.250) 0.122 (0.310) ⫺3.271*** (0.376) ⫺4.865*** (0.377) ⫺5.059*** (0.396) ⫺2.821*** (0.288) ⫺1.268*** (0.389) ⫺1.890*** (0.428)
Constant 83.604*** (0.273) 81.781*** (0.254) 83.174*** (0.298) 84.671*** (0.348) 89.823*** (0.296) 89.695*** (0.329) 80.257*** (0.271) 81.017*** (0.468)
Observations 6,967 6,147 4,161 3,425 3,675 3,645 5,924 3,366
R2 0.009 0.003 0.005 0.016 0.017 0.009 0.008 0.006

Notes: Standard errors in parentheses; *significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
 Dimension 1: Dimension 2:
Quality of Transparency Dimension 3: Dimension 4: Dimension 5: Dimension 6:
Explanatory Weighted Unweighted participation in of local Downward Control of Administrative Public service
variables PAPI in 2011 PAPI in 2011 village decisions decision-making accountability corruption procedures delivery

Districts covered
in PAPI 2010 0.453*** (0.054) 0.455*** (0.054) 0.088*** (0.015) 0.088*** (0.018) 0.079*** (0.015) ⫺0.023 (0.020) 0.117*** (0.011) 0.105*** (0.007)
Provincial People’s Committee
is located in districts 0.932*** (0.046) 0.652*** (0.046) 0.163*** (0.012) 0.231*** (0.014) 0.119*** (0.012) ⫺0.044** (0.017) ⫺0.060*** (0.010) 0.243*** (0.007)
Total population of
district in 2009 0.112 (0.203) 0.503** (0.202) ⫺0.002 (0.049) ⫺0.212*** (0.063) ⫺0.182*** (0.058) 1.295*** (0.081) 0.028 (0.048) ⫺0.423*** (0.031)
Total population of
province in 2009 ⫺0.028 (0.033) ⫺0.283*** (0.033) ⫺0.061*** (0.009) ⫺0.109*** (0.010) ⫺0.086*** (0.009) ⫺0.072*** (0.012) ⫺0.085*** (0.007) 0.129*** (0.005)
City class 1 ⫺0.258*** (0.054) ⫺0.450*** (0.056) ⫺0.050*** (0.016) ⫺0.082*** (0.019) ⫺0.033** (0.015) ⫺0.401*** (0.019) ⫺0.040*** (0.012) 0.158*** (0.007)
Constant 38.079*** (0.043) 36.933*** (0.043) 5.291*** (0.012) 5.739*** (0.014) 5.663*** (0.012) 6.287*** (0.015) 7.106*** (0.009) 6.846*** (0.006)
Observations 7,507 7,507 7,507 7,507 7,507 7,507 7,507 7,507
R2 0.017 0.012 0.007 0.011 0.006 0.014 0.006 0.051

Notes: Standard errors in parentheses; *significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey
105
monitoring
PAPI

with control variables

Regressions of the
Table AV.

covered in PAPI 2010:

2011 PAPI on districts
 IJDI
16,1

106

Table AVI.

control variables
2011 satisfaction
Regressions of the

districts covered in
PAPI 2010: without
about governance on
Satisfaction about
Commune Satisfaction about Satisfaction about Your satisfaction Your satisfaction
Satisfaction about People’s District People’s Province People’s Satisfaction about about National about commune Satisfaction about
Explanatory variables village head Committee Committee Committee Government Assembly security district court

Districts covered in PAPI

2010 3.402*** (0.279) 2.319*** (0.286) 2.046*** (0.304) 2.549*** (0.326) 2.327*** (0.343) 2.128*** (0.324) 2.548*** (0.291) 1.674*** (0.419)
Provincial People’s Committee
is located in districts 0.846*** (0.218) 0.784*** (0.221) ⫺0.828*** (0.244) ⫺2.619*** (0.265) ⫺2.359*** (0.245) ⫺1.869*** (0.240) 1.387*** (0.253) 0.405 (0.291)
Total population of district in
2009 ⫺4.013*** (1.372) ⫺5.653*** (0.838) 6.225*** (0.924) ⫺2.579** (1.060) ⫺2.973*** (0.970) 2.263** (0.990) ⫺3.779*** (1.056) 0.716 (1.613)
Total population of province
in 2009 ⫺0.362** (0.161) 0.672*** (0.171) ⫺0.227 (0.191) 1.607*** (0.220) 0.971*** (0.203) 0.608*** (0.197) 0.679*** (0.194) 1.143*** (0.228)
City class 1 0.908*** (0.244) 0.662** (0.305) ⫺2.693*** (0.370) ⫺4.100*** (0.370) ⫺4.448*** (0.386) ⫺2.189*** (0.282) ⫺0.473 (0.373) ⫺1.015** (0.404)
Constant 84.664*** (0.238) 82.248*** (0.228) 83.744*** (0.273) 85.475*** (0.315) 90.552*** (0.263) 90.440*** (0.287) 81.286*** (0.243) 82.388*** (0.399)
Observations 6,967 6,147 4,161 3,425 3,675 3,645 5,924 3,366
R2 0.008 0.004 0.005 0.015 0.017 0.008 0.006 0.003

Notes: Standard errors in parentheses; *significant at 10%; ** significant at 5%; *** significant at 1%
Source: Estimation from the 2011 PAPI survey